Starting a rebuttal with the word "wrong" seems rather counter-productive. If you have a reasonable point that disproves mine, then anybody reading it would notice this and as such the "wrong" is not required. If on the other hand, your point does not in fact counteract mine, then making such a strong statement seems rather foolish.
To begin:
* Uranite (U3O8) is a major ore of uranium https://secure.wikimedia.org/wikipedia/en/wiki/Uraninite and is in fact mostly uranium by weight.
* "Uranium mining is the largest business in Niger and the life expectations is not much about 50 years there." Is an example of 'correlation does not imply causation'. To give you an example, the number of pirates have gone down over the last few hundred years. Global temperatures have gone up over the last few hundred years. From that information alone you cannot say that the lack of pirates causes global warming.
* I'd also mention here that there are large amounts of diamonds sourced from africa. A trade that is very profitable, yet leaves the people themselves with very little money nor increased standard of living.
* From the wikipedia article you specified: "Because uranium ore emits radon gas, uranium mining can be more dangerous than other underground mining, unless adequate ventilation systems are installed.". Assuming these safeguards are met, I fail to see how it would be 2,000,000 times as dangerous as coal mining.
I will, however, correct my previous statement.
-Natural uranium has an energy density of 443,000MJ/kg
-Coal has an energy density of 32.5MJ/kg.
Hence to produce the same amount of energy as you can produce using one KG of natural uranium you require 13630kg of coal.
Assuming the risk of coal mining is measured in a certain number of fatalities per kilogram mined (equivalent to a certain m^3 mined, given a constant density of the mined material) is R_C and R_U for coal and uranium respectively then the expected number of fatalities per "1kg of uranium equivalent energy" is then R_U and 13630R_C.
As such for coal to be "safer" 13630R_C needs to be < R_U, a statement I consider unlikely.